An analysis of the spillover effects on African stock markets
The respective values of the spillover matrix are sensitive to the order of the data in In Diebold and Yilmaz (2009, Economic Journal)’s methodology. To remove this shortcoming, Diebold and Yilmaz (2012) examine the volatility spillovers by inventing a revised version of spillover measure based on the generalized impulse response approach by Koop, Pesaran and Porter (1996, Journal of Econometrics) and Pesaran and Shin (1998, Economics Letters). The forecast error variance decompositions of this version are independent to variable ordering. Accordingly, this paper uses the new method created by Diebold and Yilmaz (2012).
We find that African markets are most severely affected by spillovers from global markets and modestly from commodity and currency markets. Conversely, the regional spillovers within Africa are smaller than the global ones and are insulated from the global crises. We also find that the aggregated spillover effects of European countries to the African markets exceeded that of the US even at the wake of the US financial crisis.